Staff Forward Deployed Engineer, Afresh

• You like owning a customer's result from the first call through rollout
• You can start before the customer knows exactly what they need
• You want to see people use what you built
• We encourage all highly qualified candidates to apply, even if they don't meet every point below
• 4+ years of experience building software professionally, with real depth in at least one area: backend or full-stack engineering, data engineering, data science, or AI engineering. You need one you're strong in, and the appetite to work across the rest
• Customer-facing delivery experience: consulting, forward-deployed, solutions, or implementation work where you sat with a client, worked out what they needed, and delivered it. You're an engineer who can run the customer conversation, then go build what came out of it
• You can take a vague problem, ask good questions, and turn it into a plan someone will sign off on
• You can explain a technical tradeoff to an engineer and to a VP of store operations
• You learn unfamiliar systems quickly, whether that's a grocer's ERP, their merchandising tools, or their data warehouse
• AI-native: You use AI tools to build more effectively and have shipped an AI-enabled product or workflow. You understand where these systems are useful, how they fail, and how to evaluate their quality
• You move fast and finish what you start
• You're comfortable traveling to customer sites regularly (~10–20%)

Desirable

• Afresh's AI already runs fresh ordering and inventory for many of the largest grocers in the US. Now we're applying that foundation to more of the business—building workflows and applications that automate manual work, surface the right information at the right time, and help teams make better decisions
• As a Forward Deployed Engineer, you're the technical lead on a customer deployment. You work in a small pod with a Deployment Strategist. They own the account, the project management, and the commercial side. You own the technical side: working out with the customer what to build, designing it, building it, and getting it into production. Success is measured on the customer's outcome (hours saved, waste cut, better decisions)
• Most of your week is hands-on building, roughly three-quarters. The rest is with the customer: discovery sessions, working sessions with their engineers, and demos of what you just shipped. The best deployments come from engineers who heard the problem firsthand
• You build on a shared platform that our AI Platform team owns: the data and context layer, the agent framework, and the tooling to deploy them. You don't start from scratch at each customer. When you uncover a gap or an opportunity in the field, you help turn it into a platform improvement—sometimes by making the change yourself, sometimes in partnership with the AI Platform team. What you learn with one customer can make every deployment that follows better
• This is 0-to-1 work for engineers with a high degree of ownership and autonomy: you build ground-up solutions for customers and see them through to impact. It’s a fast-paced environment—a pilot takes a week or two, a rollout takes two to three months, and then you move on to the next customer, carrying what you learned forward
• Work with the customer's business and technical teams, up to their leadership, to turn a broad goal like "cut shrink in produce" into a scoped problem. Who will use it? What data exists? What does done look like? What won't we build?
• Find what the customer isn't telling you. That might be data that doesn't exist, a process that lives in one person's spreadsheet, or a security review that will take six weeks. Push back when what they're asking for won't get them the result they want
• Design and build the solution on Afresh's platform: data pipelines, integrations, agents, workflows, and user-facing applications
• Get a working pilot in front of the people who will use it, fast
• Get it running in the customer's environment, working with their engineers and our platform team on the infrastructure
• Watch how people actually use it, check that its answers are right, and iterate until it's part of how they work
• Bring what you learned back to the platform. Work with the platform team to decide what becomes part of the platform and what stays custom, so the next deployment needs less custom work